|
(T/MF165) Statistical power analysis of phytotoxicity data on Myriophyllum sp. from microcosm experiments. Sanderson, Hans1, Hanson, Mark2, 1 2 ABSTRACT- Considerations concerning the choice of scale, descriptive endpoints and number of replicates in order to obtain ecological relevance and statistical power are vital prior to designing an ecological risk assessment. This paper aims to quantify uncertainty aspects of these issues regarding phytotoxicity testing of submerged aquatic macrophytes in microcosms via statistical power analysis. The common macrophytes Myriophyllum spicatum and Myriophyllum sibiricum were used in three different ecological risk assessments at the microcosm facility at University of Guelph, Canada, during the summer of 1999. The triplicated microcosm ponds (12.000L) were treated with monochloroacetic acid, dichloroacetic acid and chlorodifluroacetic acid. Growth inhibition was estimated according to ASTM guideline using nine different endpoints over a periode of 21 days. With our analysis we wanted to quantify how many replicates that were needed to detect an ecologically significant effect with high power ( Key words: power analysis, microcosm, myriophyllum, haloacetic acids |
≤.05 and 
≤0.2) for each endpoint and thus to determine which endpoint that from a statistical point of view yields the most powerful reults. We found that, node number was consistenly the most powerful endpoint. It was possible to detect a 20-25% change from the controls with high power. The typical range of detectable change with three replicates with high power was 40-60%. Pigmentation and wet and dry mass endpoints yielded the least powerful results with ranges of 60-80% change needed to obtain high power with only three replicates. However, initial growth inhibition of 10-20% for Myriophyllum sp. are considered ecologically significant and would require 10-100 replicates in this design. This indicates severe problems in interpreting ecological risk assessment of phytotoxicity of haloacetic acids on submerged macrophytes in microcosms in terms of risk of committing a Type II error.